Wrapping up the HA-600A Product Detector Project -- Let's Call Them "Crossed Diode Mixers" NOT "Diode Rings"

This has been a lot of fun and very educational.   The problem I discovered in the Lafayette HA-600A product detector caused me to take a new look at how diode detectors really work.  It also spurred me to make more use of LTSpice.  

In the end, I went with a diode ring mixer. Part of this decision was just my amazement at how four diodes and a couple of transformers can manage to multiply an incoming signal by 1 and -1, and how this multiplication allows us to pull audio out of the mess. 

But another part of the decision was port isolation: the diode ring mixer with four diodes and two transformers does keep the BFO signal from making its way back to into the IF chain.  This helps prevent the BFO signal from activating the AGC circuitry, and from messing up the S-meter readings. LTSpice helped me confirm that this improvement was happening:  in LTSpice I could look at how much BFO energy was making its way back to the IF input port on the diode ring mixer.  LTSpice predicted very little, and this was confirmed in the real world circuit. (I will do another post on port isolation in simpler, singly balanced diode mixers.)  

At first I did have to overcome some problems with the diode ring circuit.  Mine seemed to perform poorly with strong signals: I'd hear some of the "simultaneous envelope and product detection" that started me down this path.  I also noticed that with the diode ring, in the AM mode the receiver seemed to be less sensitive -- it was as if the product detector circuit was loading down the AM detector.  

One of the commenters -- Christian -- suggested putting some resistance into the input of the diode ring circuit.  I put a 150 ohm pot across the input, after the blocking capacitor. The top of the pot goes to the capacitor, the bottom to ground and the wiper to the input of L1 in the diode ring circuit (you can see the circuit in the diagram above).  With this pot I could set the input level such that even the strongest input signals did not cause the envelope detection that I'd heard earlier.  Watching these input signals on the 'scope, I think these problems arose when the IF signals rose above .7 volts and started turning on the diodes.  Only the BFO signal should have been doing that.  The pot eliminated this problem.   The pot also seemed to solve the problem of the loading down of the AM detector.  

With the pot, signals sounded much better, but I thought there was still room for improvement.  I thought I could hear a bit of RF in the audio output.  Perhaps some of the 455 kHz signal was making it into the AF amplifiers.   I looked at the circuit that Wes Hayward had used after the SBL-1 that he used as product detector in his Progressive Receiver.  It was very simple:  a .01 uF cap and 50 ohm resistor to ground followed by an RF choke.  I can't be sure, but this seemed to help, and the SSB now sounds great. 

A BETTER NAME? 

One suggestion:  We should stop calling the diode ring a diode ring.  I think "crossed diode mixer" or something like that is more descriptive.  This circuit works not because the diodes are in a ring, but because two of them are "crossed."   From now on I intend to BUILD this circuit with this crossed parts placement -- this makes it easier to see how the circuit works, how it manages to multiply by -1, and to avoid putting any of the diodes in backwards.

I prefer the bottom diagram

A KNOWN PROBLEM? 

I'm left wondering if the engineers who designed the HA-600A were aware of the shortcomings of the product detector.  It is really strange that my receivers lacks a 12V line from the function switch to the product detector. And it is weirder still that the detector works (poorly) even with no power to the transistor.  What happened there?  

When you look at the HA-600A manual, you can see a hint that maybe they knew there was a problem.  For CW and SSB, the manual recommends leaving the AF control at the quarter or halfway point, then controlling loudness with the RF gain control.  This would have the effect of throttling back the RF gain (and the potential for product detector overload) when strong signals appear.  MGC in addition to the AGC.  Any memories or insights on this would be appreciated.